herbps10

Gaussian Processes in particular can be very useful in regression problems when you don't want to make strong assumptions about the functional form of the relationship between variables. (You can still introduce more general assumptions, like that the relationship is "smooth" to some degree or is periodic, by your choice of GP kernel function.)

This is great, using Brownian motion to introduce stochastic processes is a neat way to introduce the idea of priors over a function space.

My small contribution to this area is an interactive notebook where you can add data points and see how different GP kernels behave: https://observablehq.com/@herbps10/gaussian-processes

French. I use Anki for building vocabulary, and practice over Skype. The innerFrench YouTube channel has also been a great resource.

Another excellent choice for a LaTex editor is LyX (https://www.lyx.org/). It's not WYSIWYG, but does basic math formatting as you go, which is very helpful for complicated expressions.

Most emerging contaminants have no clinical reference levels for safe vs. unsafe exposure, so often the best you can do is compare your exposure level to a reference distribution to see if you have relatively high levels.

The CDC tests thousands of people in the U.S. every two years for a panel of chemicals (including endocrine disruptors) via NHANES, so at very least you can compare your results to that to see if you are unusual relative to a nationally representative sample of Americans.

I used to work on an endocrine disruptor (EDC) testing program (https://wwww.silentspring.org/detoxmeactionkit/). Analytical laboratories won't deal with individual samples from consumers, so our approach was to collect batches that we sent to the lab all at once. The cost adds up quickly because (1) the analytical chemistry done for the testing requires a mass spectrometer, which need constant expert attention and expensive reference standards, (2) urine needs to be mailed frozen, which requires overnight shipping. The program I worked on didn't test for phthalates because IIRC that would require sending the samples through the mass spec a second time, which substantially increases the price.

I think it would be possible to bring the price down if you had your own dedicated mass spec and high volume, but it's never going to be cheap. Unfortunately the usefulness of this kind of testing increases the more you can get tested, e.g. before and after a dietary intervention. I think the market for this kind of testing could be large if there was a way to bring the price down, but that probably requires the development of new analytical chemistry techniques.

Silent Spring Institute | Newton, MA | ONSITE | Software Engineer We’re looking for a web application engineer to lead our efforts to design digital tools to communicate research about harmful chemicals in everyday environments.

We measure chemical levels in people’s environments and in their bodies—even a New York Times columnist Nicholas Kristof was tested in one of our studies: http://nyti.ms/2CdcyQ4! We built a web application to help us return individual results in an understandable way to the people who were tested. We need your help turning it into a scalable tool that can be used by more studies to reach tens of thousands of participants. Here’s a description of some of our work so far: https://ehp.niehs.nih.gov/EHP702/.

Silent Spring is relentlessly asking tough questions about the chemicals in our homes and in the products we buy. You’ll be there with us, helping us communicate our findings to create conscious consumers and healthier communities.

You will have opportunities to develop new software tools or lead data science projects, for example, analyzing environmental health data, developing new R packages, or creating novel visualizations of our unique datasets. You will help develop new smartphone tools, like Detox Me (https://silentspring.org/detoxme/) to collect data on the products people are using and track interventions to their reduce exposures.

This is a full time position. We offer competitive salaries, 4 weeks vacation plus holidays, health insurance, and other benefits.

More information: https://silentspring.org/opportunities-silent-spring-institu... To apply: Send a cover letter and resume to [email protected]

Silent Spring Institute | Newton, MA | ONSITE | Software Engineer

We’re looking for a web application engineer to lead our efforts to design digital tools to communicate research about harmful chemicals in everyday environments.

We measure chemical levels in people’s environments and in their bodies—even a New York Times columnist Nicholas Kristof was tested in one of our studies: http://nyti.ms/2CdcyQ4! We built a web application to help us return individual results in an understandable way to the people who were tested. We need your help turning it into a scalable tool that can be used by more studies to reach tens of thousands of participants. Here’s a description of some of our work so far: https://ehp.niehs.nih.gov/EHP702/.

Silent Spring is relentlessly asking tough questions about the chemicals in our homes and in the products we buy. You’ll be there with us, helping us communicate our findings to create conscious consumers and healthier communities.

You will have opportunities to develop new software tools or lead data science projects, for example, analyzing environmental health data, developing new R packages, or creating novel visualizations of our unique datasets. You will help develop new smartphone tools, like Detox Me (https://silentspring.org/detoxme/) to collect data on the products people are using and track interventions to their reduce exposures.

This is a full time position. We offer competitive salaries, 4 weeks vacation plus holidays, health insurance, and other benefits.

More information: https://silentspring.org/opportunities-silent-spring-institu... To apply: Send a cover letter and resume to [email protected]

My understanding is that just because you live in a blue area of the map doesn't mean you have contaminated water.

Also, note that this paper only looks at public water supply testing data, not private well water.

From the linked article: "From the authors of the study: We have mapped watersheds in the United States that have potentially high concentrations of PFASs based on U.S. EPA data. This does not mean that all drinking water supplies within the highlighted regions contain high PFAS concentrations, but that at least one sample from at least one water supply was reported to be at or above levels considered safe by the U.S. EPA between 2013 and 2015. However, no measurements have been made in many water supplies across the country. We recommend increased monitoring of these contaminants in our drinking water. For more information, please contact the EPA: Cathy Milbourn, [email protected], or Monica Lee, [email protected]."

I'll follow up with my colleague to see if she has any thoughts.

The best way to prevent this, in my opinion, is to be clear that the research has nothing to do with fluoridated water.

From my colleague: "The highly fluorinated chemicals that we discussed are organic compounds, meaning that they contain carbon atoms. They are called highly fluorinated or per- and polyfluoro alkyl substances because they have many fluorine atoms, and the carbon atoms are entirely or often bound to fluorine (There may be highly fluorinated chemicals that aren't strictly PFASs, but that's a minor point). These highly fluorinated chemicals have unique chemical properties that give them their characteristic grease and water resistance (typically one end of the molecule is more soluble in oil, and the other end is more soluble in water, similar to other surfactants). The concerns about PFAS toxicity stem from the properties of the whole molecule and not just because they contain fluorine."

"The contamination of drinking water in Hoosick Falls is an example of the shortcomings of the UCMR3 program. As part of the UCMR program, all large public water supplies that serve over 10,000 residents, and a small percent of smaller public water supplies, are required to conduct water testing. Most smaller public water supplies and private wells are not required to conduct water testing as part of the UCMR program. The public water supply in Hoosick Falls serves less than 10,000 residents, so it was not required to conduct testing for PFOA as part of the UCMR program, nor were residents in the area who relied on private wells."

I can add that we're also following several other communities with PFAS contamination issues, like Newburgh, NY; Bennington, VT; and Pease, NH.

I work with one of the authors of the Hu et al. paper that this article is based on. I'm happy to relay any questions people have to her.

“What are the chances of falling off the roof?” he argued vociferously when we tried previously to hash it out. Have I ever known anyone who has fallen off a roof? Anyway, he said, it’s not as if he doesn’t give his kids any limits: They are not allowed to play ball or tag up there."

Yes, but most kids are not allowed or encouraged to play on their roofs. If they were, the rate of injury would likely be much higher.

This is great to see as I've been working on a similar project to try to visualize relationships by looking at the number of texts sent over time.

If anyone would like help generating similar analyses of their texting data, I'd be glad to help as I have some machinery set up to do so!

Here's a prototype site I put together that takes iPhone SMS backups and generates a graph of how many texts you've sent over time:

http://herbsusmann.com/relationships/